We compare prompted and fine-tuned language models for verifying numerical claims in the CheckThat! Lab Task 3 challenge, exploring retrieval and model adaptation strategies to improve numerical claim verification performance.
Anirban Saha Anik, Md Fahimul Kabir Chowdhury, Andrew Wyckoff, Sagnik Ray Choudhury. "ClaimIQ at CheckThat! 2025: Comparing Prompted and Fine-Tuned Language Models for Verifying Numerical Claims." Working Notes of the Conference and Labs of the Evaluation Forum (CLEF 2025).
We introduce a dynamic fusion framework that combines multiple LLMs and retrieval to produce consistent, professional, and actionable crisis responses, and propose a consistency metric to measure stylistic stability across outputs.
Song, Xiaoying, Anirban Saha Anik, Eduardo Blanco, Vanessa Frias-Martinez, and Lingzi Hong. "A Dynamic Fusion Model for Consistent Crisis Response." Findings of the Association for Computational Linguistics: EMNLP 2025.
We propose a multi-agent retrieval-augmented framework that orchestrates specialized LLM agents for evidence retrieval, summarization, generation, and refinement to produce evidence-based counterspeech that outperforms RAG baselines.
Anirban Saha Anik, Xiaoying Song, Elliott Wang, Bryan Wang, Bengisu Yarimbas, Lingzi Hong. (2025). "Multi-Agent Retrieval-Augmented Framework for Evidence-Based Counterspeech Against Health Misinformation." Conference on Language Models (COLM 2025).
We propose a dynamic fusion approach that integrates multiple LLMs and retrieval to generate timely, professional, and actionable crisis responses for social media contexts.
Hong, L., Song, X., Saha Anik, A., Frias-Martinez, V. (2025). "Dynamic Fusion of Large Language Models for Crisis Communication." Proceedings of the International ISCRAM Conference.
Authors: Mithila Arman, Naheyan Prottush, Maher Ali Rusho, Arup Datta, Anirban Saha Anik, Din Mohammad Dohan, Md. Ashiq Ul Islam Sajid, Intezab Alam Sheikh, Md. Khurshid Jahan
Conference:2025 IEEE 4th International Conference on Computing and Machine Intelligence (ICMI)
We introduce a hybrid attention-guided fusion network with Grad-CAM for interpretable and accurate MPox skin lesion classification, demonstrating high diagnostic performance on a curated dataset.
M. Arman, N. Prottush, M. A. Rusho, A. Datta, Anirban Saha Anik, D. M. Dohan, M. A. U. I. Sajid, I. A. Sheikh, and M. K. Jahna. "A Hybrid Attention-Guided Fusion Network with Grad-CAM for MPox Skin Lesion Classification." 2025 IEEE 4th International Conference on Computing and Machine Intelligence (ICMI).
We apply transformer-based NLP to analyze student feedback and sentiment to assess Outcome-Based Education (OBE) effectiveness and surface actionable insights for improving educational outcomes.
Das, Shuvra Smaran and Anik, Anirban Saha and Morol, Md Kishor and Mahmood, Mohammad Sakib. "Outcome-Based Education: Evaluating Students Perspectives Using Transformer." 2024 27th International Conference on Computer and Information Technology (ICCIT).
We use modified artificial neural networks to predict future lockdown timings based on epidemiological and socio-economic features, achieving high accuracy on a curated pandemic dataset.
Shuvra Smaran Das, Anirban Saha Anik, Md. Muzakker Hossain, Md. Kishor Morol, Fariha Jahan, Md. Abdullah Al-Jubair. "A Study on Future Lockdown Predictions Using ANN." International Conference on Next-Generation Computing, IoT and Machine Learning (NCIM 2023). DOI: 10.1109/NCIM59001.2023.10212686.
We compare multiple deep CNN architectures on a large chest X-ray dataset (COVIDx CXR-3) for COVID-19 detection, finding EfficientNet-B3 yields the best balance of accuracy and sensitivity.
Anirban Saha Anik, Kowshik Chakraborty, Bishowjit Datta, Abdul Kader, MD. Kishor Morol. "A Comparative Analysis for the Detection of COVID-19 from Chest X-ray Dataset." International Conference on Recent Progresses in Science, Engineering and Technology (ICRPSET 2022). DOI: 10.1109/ICRPSET57982.2022.10188570.
